Designing Ontology-Based Data Access Solutions: Representation and Reasoning Support

While OBDA has been the focus of many different studies in recent years, these studies typically have focused on the core aspects of OBDA, in particular on query answering. Less attention, in this context, has been dedicated to other aspects such as ontology representation, ontology visualization, and intentional reasoning. From the experience matured in the last few years in working with OBDA, it is clear that these problems, while also representing independent research fields, all play a significant part in the construction of the architecture of a full-fledged OBDA system, and therefore must be addressed. Unfortunately, the results that are found in literature regarding these issues are unsatisfactory when taken in the context of OBDA, because they are not tailored towards the languages and the applications that are adopted under the OBDA framework. This thesis addresses these issues and presents solutions for them in the context of the OBDA framework. The integration of these solutions and techniques into an OBDA system allows to provide users with full support for all OBDA-related activities. We briefly summarize in the following the main contributions provided in this thesis: we present the novel Graphol language for the graphical specification of DL ontologies; we study the problem of approximating DL ontologies specified in an expressive source language in terms of less expressive target languages; we deal with core intentional reasoning services over an ontology, devising efficient methods and implementations that are specifically tailored towards languages that are suitable for OBDA.